C
2014. The American Astronomical Society. All rights reserved. Printed in the U.S.A.
HERSCHEL
EVIDENCE FOR DISK FLATTENING OR GAS DEPLETION IN TRANSITIONAL DISKS
∗J. T. Keane1, I. Pascucci1, C. Espaillat2, P. Woitke3, S. Andrews4, I. Kamp5, W.-F. Thi6, G. Meeus7, and W. R. F. Dent8
1Lunar and Planetary Laboratory, University of Arizona, Tucson, AZ 85721, USA 2Department of Astronomy, Boston University, Boston, MA 02215, USA
3SUPA, School of Physics & Astronomy, University of St. Andrews, North Haugh, St. Andrews KY16 9SS, UK 4Harvard-Smithsonian Center for Astrophysics, Cambridge, MA 02138, USA
5Kapteyn Astronomical Institute, Postbus 800, 9700 AV Groningen, The Netherlands
6Institut de Plan´etologie et d’Astrophysique (IPAG) UMR 5274, Universit´e Joseph Fourier Grenoble-1, CNRS-INSU, F-38041 Grenoble, France 7Departamento de F´ısica Te´orica, Universidad Aut´onoma de Madrid, Campus Cantoblanco, E-28049 Madrid, Spain
8ALMA SCO, Alonso de Cordova 3107, Vitacura, Santiago, Chile
Received 2014 January 8; accepted 2014 March 28; published 2014 May 14
ABSTRACT
Transitional disks are protoplanetary disks characterized by reduced near- and mid-infrared emission, with respect to full disks. This characteristic spectral energy distribution indicates the presence of an optically thin inner cavity
within the dust disk believed to mark the disappearance of the primordial massive disk. We present newHerschel
Space ObservatoryPACS spectra of [Oi] 63.18μm for 21 transitional disks. Our survey complements the larger
Herschel GASPS program (“Gas in Protoplanetary Systems”) by quadrupling the number of transitional disks
observed with PACS in this wavelength. [Oi] 63.18μm traces material in the outer regions of the disk, beyond the
inner cavity of most transitional disks. We find that transitional disks have [Oi] 63.18μm line luminosities∼2 times
fainter than their full disk counterparts. We self-consistently determine various stellar properties (e.g., bolometric
luminosity, FUV excess, etc.) and disk properties (e.g., disk dust mass, etc.) that could influence the [Oi] 63.18μm
line luminosity, and we find no correlations that can explain the lower [Oi] 63.18μm line luminosities in transitional
disks. Using a grid of thermo-chemical protoplanetary disk models, we conclude that either transitional disks are less flared than full disks or they possess lower gas-to-dust ratios due to a depletion of gas mass. This result suggests that transitional disks are more evolved than their full disk counterparts, possibly even at large radii.
Key words: accretion, accretion disks – circumstellar matter – infrared: stars – protoplanetary disks – stars: pre-main sequence
Online-only material:color figures, extended figures
1. INTRODUCTION
Protoplanetary disks (gas-rich dust disks around young stars) provide the raw building blocks for solar systems. While significant progress has been made in understanding the relevant evolutionary timescales of protoplanetary disks, little is known about the physical mechanisms driving the eventual dispersal of dust and gas about these young systems (for review, see
Pascucci & Tachibana2010). The goal of this paper is to gain
insight into these dispersal processes by investigating a special type of protoplanetary disk that is thought to be in the process of losing its primordial dust disk: the transitional disks.
Transitional disks, like their full protoplanetary disk cousins, are often identified by their spectral energy distributions (SEDs). While there is significant variation in the SEDs of young star systems, transitional disks appear as a distinct subgroup of protoplanetary disks: their SEDs show reduced near- and
mid-infrared emission, with respect to full disks (Strom et al.1989).
This characteristic SED points to the presence of an optically
thin inner cavity, extending from the star out to 1∼20 AU. The
excavation of this cavity is believed to mark the early stages of the dispersal of the primordial, massive dust disk, whose continuous dust disk extended as close as a few stellar radii to
the central star (e.g., Calvet et al.2002; Espaillat et al.2007).
The existence of inner cavities has been directly confirmed for a few transitional disks via sensitive, high-resolution millimeter
∗ Herschelis an ESA space observatory with science instruments provided by
European-led Principal Investigator consortia and with important participation from NASA.
observations which detect reduced (or absent) dust emission from the inner disk as a result of a deficit of millimeter size
grains (e.g., Andrews et al. 2009; Brown et al.2009). While
transitional disks may possess dust cavities, it is known that, in most cases, these dust cavities are not devoid of gas. Transitional
disks are still actively accreting (e.g., Najita et al.2007), and
various optical emission lines (e.g., CO lines, [Oi] 6300 Å and
5577 Å, etc.) indicate the presence of gas within the dust cavity region, though it may be depleted (e.g., TW Hya; Gorti et al. 2011).
There are three leading hypotheses for the driving mechanism behind the formation of cavities in transitional disks (for review,
see Espaillat et al.2014).
1. Dust coagulation.As disks evolve, submicron-sized dust grains coagulate into larger aggregates which have little emission at infrared wavelengths and thus reduce the disk opacity. These larger aggregates would eventually coalesce into planetesimals and planetary embryos. Since dynamical timescales increase with increasing radial distance from the central star, grain growth occurs inside-out and leads to the development of an expanding optically thin inner cavity, although the total mass of this inner disk region is not
necessarily lower (e.g., Dullemond & Dominik2005).
2. Photoevaporation. High-energy photons from the central star can drive photoevaporative winds, particularly from the
outer regions of the protoplanetary disk (beyond∼few AU).
cavity (e.g., Alexander et al. 2014). Direct irradiation of the cavity wall is expected to rapidly disperse the outer
disk (Alexander et al.2006). Photoevaporative winds have
been detected for select protoplanetary disks via blueshifted
(∼few km s−1) [Neii] 12.81μm lines, which traces
un-bound winds within the inner10’s of AU (Pascucci &
Sterzik2009).
3. Dynamical clearing by giant planets.Dynamical interac-tions between the disk and an embedded giant planet (with masses roughly equal to that of Jupiter) can open gaps
within the disk (e.g., Lubow et al.1999). Gas from the
in-ner disk (within the planet’s orbit) can continue to accrete onto the central star, while most of the gas from the outer disk (beyond the planet’s orbit) accretes onto the planet, and only a small amount of gas flows past the planet into the inner disk. In addition to the physical gap created by the planet, pressure gradients setup at the outer edge of the gap can act as a dust filter, allowing only grains below a critical size to reach the inner disk and perhaps forming an
optically thin inner cavity (Rice et al.2006).
While these different mechanisms can produce qualitatively similar SEDs, they predict distinctive differences in the distri-bution of disk gas. Furthermore, these different processes can, and probably do, operate simultaneously.
In this paper, we useHerschel Space Observatoryfar-infrared
data to examine whether full disks and transitional disks are dif-ferent in their outer disk regions, beyond 10’s of AU. We use the
[Oi] 63.18μm emission line and the nearby 63μm continuum
emission to trace the gas and dust components, respectively,
beyond 10 ∼ 100 AU (e.g., Aresu et al. 2012). In addition,
we use ancillary data to characterize our sample at different
wavelengths. In Section2, we provide a short description of our
sample, theHerschel/PACS observations and data reduction,
and the ancillary stellar and disk properties used to characterize
our sample. In Section3, we summarize our [Oi] 63.18μm line
63μm continuum results. Most notably, we find that transitional
disks possess [Oi] 63.18μm line luminosities a factor of 2∼
3 lower than full disks, despite having similar 63μm
contin-uum luminosities, a trend previously identified by Howard et al.
(2013), though expanded in this work with quadruple the number
of transitional disks. In Section4, we rule out various observable
stellar and disk properties (e.g., FUV and X-ray luminosity) as
the potential cause for this [Oi] 63.18μm line luminosity
dif-ference between full disks and transitional disks. In Section5,
we use the results of the DENT grid (a grid of 300,000
thermo-chemical protoplanetary disk models, by Woitke et al.2010),
to examine other possible causes for the [Oi] 63.18μm line
lu-minosity difference. We conclude that the lower [Oi] 63.18μm
line luminosity of transitional disks could be due to transitional disks either being less flared or by having lower gas-to-dust
ra-tios. In Section6, we discuss the implications of this result for
disk evolution models and potential follow-up observations.
2. OBSERVATIONS AND DATA REDUCTION
2.1. Sample Description
We selected 21 transitional disks from predominantly young
(a few Myr old) and nearby (200 pc) star-forming regions.
Five additional transitional disks were selected from the GASPS
sample (“Gas in Protoplanetary Systems”; Dent et al. 2013),
resulting in a total of 26 transitional disks. Our sample is listed
in Table1. The transitional disks were identified by significant
dips in theirSpitzer/IRS spectra. (For the relevant spectra used
to identify each transitional disk as transitional, see Brown et al.
2007; Calvet et al. 2002; Espaillat et al. 2010; Furlan et al.
2009; Kim et al.2009; Mer´ın et al.2008,2010.) Because only
10% of protoplanetary disks are transitional (e.g., Williams &
Cieza 2011; Muzerolle et al. 2010), we selected targets from
a number of star-forming regions, including Taurus-Auriga, Ophiuchus, Chameleon, and Lupus. Targets that have been previously modeled either with continuum radiative transfer codes or simple prescriptions for the disk inner cavity were given preference, as were objects with archival measurements of accretion rates and infrared and millimeter observations.
For comparison with our sample of transitional disks, we selected an additional 33 protoplanetary disks from the GASPS
(“Gas Survey of Protoplanetary Systems”; Dent et al. 2013)
survey of the Taurus-Auriga star-forming region.9 These disks
were selected as being typical protoplanetary disks, with IRS
spectra close to the Taurus-Auriga mean (D’Alessio et al.2006),
and were also selected to sample similar spectral types to those of the transitional disks. Like the transitional disk subsample, we took preference for objects with known accretion rates and millimeter observations. Of these 33 protoplanetary disks, 15
have jets/outflows as identified by a combination of optical
and near-IR spectroscopy and imaging (see Kenyon et al.2008
and references therein). We will refer to this subsample as “outflow” sources. The remaining 18 disks without outflows will be referred to as “full” disks. This distinction between outflow
disks and full disks varies slightly from Howard et al. (2013),
who identified outflow disks as objects with either directly
imaged jets in Hα, [Oi]λ6300, [Sii]λ6371, being associated
with Herbig–Haro objects, or having very broad [Oi] λ6300
emission line profiles. This slight difference in definition only changes the classification of two disks (AA Tau and DL Tau). None of our full disks were noted to have broadened or spatially
extended [Oi] 63.18μm emission in Howard et al. (2013).
Targets with binary companions represent a possible source of contamination within our samples given the large spaxel
size of PACS (9.4 on a side, which corresponds to a projected
separation of∼1300 AU at the distance of Taurus). Close binary
companions can interact with the primary star’s disk and produce
transitional SEDs (e.g., CoKu Tau/4; Ireland & Kraus2008).
Medium and large separation binaries (projected separations
>10 AU) do not seem to strongly affect the first steps of
planet formation (grain growth and dust settling; Pascucci et al.
2008). However, binaries with separations40 AU significantly
hasten the process of disk dispersal (Kraus et al.2012). While
it might be best to remove all binaries from our sample and just focus on single stars, this approach could bias our samples and, as such, our results. The Taurus-Auriga star-forming region, from which we draw most of the full disks, has been well surveyed for multiplicity. However, Ophiuchus, Chameleon, and the Lupus star-forming regions, from which we draw our sample of transitional disks, have not been as well studied. Thus, there are very likely undetected multiple systems within our transitional disk sample. We opted to retain full disks within
multiple systems, as long as the mid-infrared flux (∼10μm)
ratio between members is large (LIR,primary/LIR,secondary 3),
and the protoplanetary disks are not circumbinary. Targets that do not meet this criterion are excluded from all analysis (though
they are included in tables and figures, for reference). Table1
lists the multiplicity status for all targets, as well as the relevant
9 An additional full disk, SZ 50, from the Cha I star-forming region, was
Table 1
Herschel/PACS Sample and Observations
ID Name R.A. Decl. Association SpTy Ref. Multiplicity Ref. ObsID Duration
(s) Transition Disks
T1 16201-2410F-1* 16 23 09.23 −24 17 04.70 Ophiuchus G0 F09 1 . . . 1342250127 8212
T2 CHX 22* 11 12 42 69 −77 22 23.00 Chameleon G8 L07 2 D13 1342233474 8212
T3 CHX 7* 11 06 15 41 −77 21 56.90 Chameleon G5 L07 . . . . . . 1342233477 8212
T4 CR Cha 10 59 06 99 −77 01 40.40 Chameleon K2 E11 2 G07 1342232614 8212
T5 CS Cha* 11 02 24 91 −77 33 35.70 Chameleon K6 E11 . . . . . . 1342233480 8212
T6 DM Tau 04 33 48.72 +18 10 09.99 Taurus M1 KH95 . . . . . . 1342225825 6628
T7 DoAr 28 16 26 47.42 −23 14 52.20 Ophiuchus K5 M92 . . . . . . 1342241707 8212
T8 DoAr 44 16 31 33.46 −24 27 37.30 Ophiuchus K3 M92 . . . . . . 1342250578 8212
T9 GM Aur 04 55 10.99 +30 21 59.25 Taurus K5.5 E11 . . . . . . 1342191357 1252
T10 Hn 24* 13 04 55 75 −77 39 49.50 Chameleon M0.5 M10 2 B96 1342235656 8212
T11 LkCa 15 04 39 17.80 +22 21 03.48 Taurus K5 KH95 . . . . . . 1342225798 6628
T12 LkHalpha 330* 3 45 48 28 +32 24 11.90 Perseus G3 BR07 . . . . . . 1342238377 8212
T13 RXJ1615.3-3255 16 15 20 23 −32 55 05.10 Lupus K4 M10 . . . . . . 1342229825 8212
T14 SSTLup 16 10 29.60 −39 22 15.00 Lupus M5 M10 . . . . . . 1342241709 8212
T15 Sz 111 16 08 54 69 −39 37 43.10 Lupus M1.5 H94 . . . . . . 1342220928 8212
T16 Sz 18 11 07 19 15 −76 03 04.80 Chameleon M2.5 L07 . . . . . . 1342232585 8212
T17 Sz 27 11 08 39 05 −77 16 04.20 Chameleon K8 L07 . . . . . . 1342233476 8212
T18 Sz 45 11 17 37 01 −77 04 38.10 Chameleon M0.5 L07 . . . . . . 1342233475 8212
T19 Sz 84 15 58 02 53 −37 36 02.70 Lupus M5.5 M10 . . . . . . 1342229826 8212
T20 Sz 91 16 07 11 61 −39 03 47.10 Lupus M0.5 H94 . . . . . . 1342229827 8212
T21 Sz Cha 10 58 16 77 −77 17 17.10 Chameleon K0 E11 2 D13 1342233478 8212
T22 T Cha* 11 57 13 53 −79 21 31.50 Chameleon G8 BR07 . . . . . . 1342232294 2068
T23 TW Hya 11 01 52 03 −34 42 18.60 TW Hydra K6 R06 . . . . . . 1342187127 1252
T24 UX Tau 04 30 03.76 +18 13 49.88 Taurus K2 KH95 3 M06 1342214357 1252
T25 WSB60 16 28 16.51 −24 36 58.00 Ophiuchus M4.5 WMRG05 . . . . . . 1342250128 8212
T26 YLW8* 16 27 10 28 −24 19 12.70 Ophiuchus G2.5 BR07 2 M06 1342229824 2068
Full Disks
F1 AA Tau 04 34 55.42 +24 28 53.16 Taurus K7 KH95 . . . . . . 1342225758 6628
F2 BP Tau* 04 19 15.84 +29 06 26.94 Taurus K7 KH95 . . . . . . 1342225728 3316
F3 CI Tau 04 33 52.00 +22 50 30.18 Taurus K7 KH95 . . . . . . 1342192125 1252
F4 CY Tau 04 17 33.73 +28 20 46.85 Taurus M1 KH95 . . . . . . 1342192794 1252
F5 DE Tau 04 21 55.64 +27 55 06.06 Taurus M2 KH95 . . . . . . 1342192797 1252
F6 DK Tau 12 53 17.23 −77 07 10.70 Taurus K7 KH95 2 WG01 1342225732 3316
F7 DL Tau 04 33 39.06 +25 20 38.23 Taurus K7 KH95 . . . . . . 1342225800 6628
F8 DN Tau 04 35 27.37 +24 14 58.94 Taurus M0 KH95 . . . . . . 1342225757 3316
F9 DQ Tau* 04 46 53.04 +17 00 00.50 Taurus M0 KH95 2 AW05 1342225806 1252
F10 DS Tau 04 47 48.21 29 25 13.83 Taurus K5 KH95 2 AW05 1342225851 3316
F11 GG Tau* 04 32 30.35 +17 31 40.60 Taurus K7 KH95 2 WG01 1342192121 1252
F12 GO Tau 04 43 03.10 +25 20 18.75 Taurus M0 KH95 . . . . . . 1342225826 3316
F13 HBC 347 03 29 38.24 +24 30 37.74 Taurus . . . . . . . . . . . . 1342192136 1252
F14 HK Tau 04 31 50.67 +24 24 17.44 Taurus M0.5 KH95 2 WG01 1342225736 3316
F15 IQ Tau 04 29 51.56 +26 06 44.89 Taurus M0.5 KH95 . . . . . . 1342225733 3316
F16 SU Aur 04 55 59.38 +30 34 01.56 Taurus G2 KH95 . . . . . . 1342217844 3316
F17 SZ 50 13 00 55.36 −77 10 22.10 Chameleon M3 HH92 . . . . . . 1342226008 3316
F18 V836 Tau 05 03 06.60 +25 23 19.71 Taurus K7 KH95 . . . . . . 1342227634 3316
Outflow Disks
O1 CW Tau 04 14 17.00 +28 10 57.83 Taurus K3 KH95 . . . . . . 1342216221 1252
O2 DF Tau* 04 27 02.80 +25 42 22.30 Taurus M3 KH95 2 P08 1342190359 1252
O3 DG Tau 04 27 04.698 +26 06 16.31 Taurus K6 KH95 . . . . . . 1342190382 1252
O4 DG Tau B 04 27 02.41 +26 05 31.76 Taurus M0 KH95 . . . . . . 1342192798 1252
O5 DO Tau 04 38 28.58 +26 10 49.44 Taurus M0 KH95 . . . . . . 1342190385 1252
O6 DP Tau 04 42 37.56 +25 15 39.62 Taurus M0.5 KH95 . . . . . . 1342191362 1252
O7 GI/GK Tau 04 55 10.85 +30 22 01.69 Taurus K6 KH95 2 AW05 1342225760 1252
O8 Haro 6-13 04 32 15.41 +24 28 59.75 Taurus M0 RM12 . . . . . . 1342192128 1252
O9 HN Tau 04 33 39.44 +17 51 52.24 Taurus K5 KH95 2 WG01 1342225796 3316
O10 HV Tau 04 38 35.38 +26 10 37.80 Taurus M1 KH95 2 WG01 1342225801 3316
O11 RW Aur 05 07 49.41 +30 24 07.65 Taurus K3 KH95 2 WG01 1342191359 1252
O12 RY Tau 04 21 57.40 +28 26 35.54 Taurus K1 KH95 . . . . . . 1342190361 1252
O13 T Tau 04 21 59.30 +19 32 08.53 Taurus K0 KH95 2 WG01 1342190353 1252
O14 UY Aur * 04 51 47.15 +30 47 14.44 Taurus K7 KH95 2 WG01 1342215699 1252
Table 1 (Continued)
Notes.Targets tagged with an asterisk were excluded from statistical tests due to either being a binary that does not meet the criteria in Section2.1, or having a spectral type earlier than K-type. BP Tau was also excluded from statistical tests, due to its nature as an “evolved” full disk. ObsIDs tagged with a star () were observed by the GASPS team and were previously reported in Howard et al. (2013), Meeus et al. (2012), and Podio et al. (2012), although they were re-reduced here using an updated version of the Herschel pipeline. Distances for each star-forming association (from Reipurth2008a,2008b): Chameleon, 160 pc; Lupus, 155 pc; Ophiuchus, 120 pc; Perseus, 250 pc; Taurus, 140 pc; TW Hya, 56 pc.
References.Andrews & Williams2005(AW05); Brandner et al.1996(B96); Brown et al.2007(BR07); Daemgen et al.2013(D13); Espaillat et al.2011(E11); Furlan et al.2009(F09); Guenther et al.2007(G07); Hughes et al.1994(H94); Kenyon & Hartmann1995(KH95); Luhman2007(L07); Magazzu et al.1992(M92); McCabe et al.2006(M06); Mer´ın et al.2010(M10); Pascucci et al.2008(P08); Riaz et al.2006(R06); White & Ghez2001(WG01); Wilking et al.2005(WMRG05).
references for the projected separations and mid-infrared flux ratio for binaries.
2.2. Herschel PACS Spectroscopy
We obtainedHerschel Space ObservatoryPACS (Poglitsch
et al. 2010) spectroscopy for our sample of 21 transitional
disks. The relevantHerschelobservation identification numbers
(ObsIDs), exposure times, and dates of our observations are
listed in Table1. The five additional transitional disks (DM Tau,
LkCa 15, GM Aur, TW Hya, and UX Tau) and the entire sample of full disks and outflow disks were previously observed as part
of theHerschelKey Program: GASPS (PI, W. Dent). We used
the line spectroscopy mode (“PacsLineSpec”) to take spectra
centered on the [Oi] 63μm line, between 62.93 and 63.43μm.
All of the observations were executed in “ChopNod” mode, in order to remove telescope emission and background.
We reduced our original observations and re-reduced the
GASPS archival data with the Herschel Interactive
Process-ing Environment (HIPE; Ott2010) version 9.0.0. We used the
default “ChopNodLineScan” pipeline along with the most re-cent calibration tree (CalTree 32). The data reduction process included: removal of saturated and overly-noisy pixels; differ-encing the on-source and off-source observations; spectral re-sponse function division; rebinning to the native resolution of
the instrument (oversample=2, upsample=1); spectral flat
fielding; and averaging over the two nod positions. We extracted our spectrum from the central spaxel and accounted for diffrac-tion losses to neighboring spaxels with an aperture correcdiffrac-tion provided in HIPE. Since outflow targets generally can have
ex-tended emission spanning multiple spaxels (Podio et al.2012;
Howard et al.2013), our measured fluxes for these outflow
tar-gets will generally underestimate their true fluxes. We verified that for all of our transitional disks and full disks, there was no appreciable emission in neighboring spaxels. This lack of extended emission also suggests that there is no significant mis-pointing in the observations of our transitional disks and full disks.
2.3. Stellar and Disk Properties
To interpret our Herschel PACS observations of [Oi] and
its relationship to the protoplanetary disk environment, we aggregated stellar properties (effective temperature, bolometric luminosity, FUV, and X-ray luminosities, etc.) as well as disk properties (disk mass, disk structure, accretion rates) that,
through past work, are known to affect the [Oi] 63.18μm
emission. In this section, we explain the methods by which we derived these stellar and disk properties. We will relate these
to ourHerschelobservations in Section4.
2.3.1. Effective Temperature and Bolometric Luminosity
We determined stellar effective temperatures by relating the host star’s spectral type (from the literature) to the corresponding
effective temperature (Luhman 1999), as listed in Table 2.
Generally, we do not assume that the effective temperatures
are accurate to more than one spectral subtype (∼100 K).
We self-consistently derived bolometric luminosities for all targets by performing a bolometric correction on de-reddened,
literature-availableI-band photometry listed in Table2.I-band
photometry is preferential, as it is less affected by intervening
dust. We de-reddened all of our I-band fluxes by relating
V-band extinctions (which are more commonly reported in the
literature) toI-band extinctions using relationships from Mathis
(1990), assumingRV values typical of the interstellar medium
(RV =3.1). The de-reddened continuum fluxes were converted
to luminosities using the known distances to each different
star-forming region (see Table 1 and references therein). Finally,
bolometric corrections from Luhman (1999) were used to
calculate bolometric luminosities for each target. These effective
temperatures and bolometric luminosities are listed in Table3.
2.3.2. FUV Luminosities and Accretion Rates
While ultraviolet observations of T Tauri stars would pro-vide the most direct measurement of the FUV luminosity, these observations are notoriously difficult. Instead, we made use of the well-known correlation between accretion rate and FUV ex-cess emission to derive FUV luminosities from stellar accretion
rates (e.g., Dahm2008; Herczeg & Hillenbrand2008).
Accre-tion rates can be determined from a large number of other more
commonly measured emission lines (e.g., Rigliaco et al.2011).
We used the Hαemission line at 6563 Å to estimate stellar
accretion rates. Hαis advantageous because it is a very
com-monly reported observational diagnostic, and it correlates well with accretion luminosities (the excess luminosity arising from the infall and accretion of material onto the central star), as
de-rived from other accretion tracers (e.g., Rigliaco et al.2011).
To determine accretion luminosities, we calculated the Hαline
fluxes from literature Hαequivalent widths (listed in Table2).
For targets with multiple Hα equivalent widths available in
the literature, we used the mean equivalent width.10 To
de-termine Hαline fluxes, we combined the Hαequivalent width
with the nearest available photometric point in the literature:
R-band. De-reddening was done similarly as for our
bolomet-ric luminosity analysis: usingV-band extinctions converted to
10 While Hαis known to be variable, it has been shown that the variability
Table 2 Literature Data
ID Rmag Ref. Imag Ref. AV Ref. HαEW Ref. logLX Ref.
(Å) (L)
Transition Disks
T1 14.20 C03 12.88 DENIS 6.90 MFM10 . . . . . . . . . . . .
T2 10.45 GS92 10.02 GS92 1.21 GS92 1.5 GS92 −3.45 FK89
T3 10.28 GS92 7.64 GS92 3.39 GS92 1.2 B08 −3.02 FK89
T4 10.46 GS92 9.73 GS92 1.50 E11 38.1 GE97 −2.94 FK89
T5 10.92 GS92 9.11 GS92 0.85 GS92 13.3 GS92 −3.36 FK89
T6 12.92 KH95 11.77 KH95 0.00 KH95 138.7 CK79 −4.33 G07
T7 12.10 C03 11.69 DENIS 2.10 CMLW95 36 M92 . . . . . .
T8 11.70 BA92 10.80 BA92 2.20 A11 68.3 BA92 −3.69 A11
T9 11.20 B93 10.70 B93 0.14 KH95 96.5 109 71 CK79,E94,C90 −3.89 A11
T10 13.00 C03 11.95 S07 2.00 M10 0.2 M10 −3.02 A00
T11 11.58 KH95 10.79 KH95 0.62 KH95 18.05 SB09 −3.99 A11
T12 11.20 C03 10.80 F95 1.55 OB95 16 SB09 . . . . . .
T13 11.28 M10 10.54 M10 1.00 M10 26 M10 −3.19 A11
T14 15.79 M10 13.90 M10 1.00 M10 18 M10 . . . . . .
T15 13.34 M08 12.17 M03 0.10 H94 145.2 H94 . . . . . .
T16 14.15 GS92 12.69 GS92 1.60 E11 5 L04 . . . . . .
T17 14.96 GS92 13.41 GS92 3.50 E11 100 L04 −4.79 W00
T18 12.57 GS92 11.59 GS92 0.60 E11 56 L04 −3.69 W00
T19 14.53 M10 12.94 M10 0.50 M10 44 M10 . . . . . .
T20 14.28 H94 12.92 H94 2.00 H94 95.9 H94 . . . . . .
T21 11.21 GS92 9.25 GS92 1.88 GS92 12 GS92 −3.99 F93
T22 11.07 S09 10.28 DENIS 1.70 S09 7.8 S09 . . . . . .
T23 11.40 DENIS 9.38 DENIS 1.00 K99 213.8 R06 −3.85 H07
T24 10.48 KH95 9.75 KH95 0.21 KH95 3.9 T09 −3.33 D95
T25 16.55 WMRG05 14.33 DENIS 2.00 WMRG05 81 WMRG05 . . . . . .
T26 12.20 DENIS 11.29 DENIS 9.00 PGS03 4 SB09 −3.59 A11
Full Disks
F1 12.06 KH95 10.99 HHG94 0.49 KH95 37.1 80 21 CK79,E94,C90 −3.49 G07
F2 11.31 KH95 10.45 KH95 0.49 KH95 40.1 55 47 49.4 CK79,E94,C90, MCH01 −3.45 G07
F3 12.22 KH95 11.12 KH95 1.77 KH95 102.1 64 CK79,C90 −4.30 G07
F4 12.35 KH95 11.18 KH95 0.10 KH95 69.5 CK79 −4.46 G07
F5 11.66 KH95 10.75 HHG94 0.59 KH95 54 76 CK79,C90 −3.78 D95
F6 11.43 KH95 10.46 KH95 0.76 KH95 19.4 13 28 CK79,E94,C90 −3.62 G07
F7 11.85 KH95 10.89 KH95 1.70 HEG95 105 111 138 WG01,CK79 C90 −3.59 D95
F8 11.49 KH95 10.49 KH95 0.49 KH95 11.9 22 15 11.1 CK79,E94,C90, MCH01 −3.52 G07
F9 12.40 KH95 11.27 KH95 0.97 KH95 112.9 CK79 . . . . . .
F10 11.56 KH95 10.80 KH95 0.31 KH95 38.5 K98 . . . . . .
F11 11.31 WG01 10.44 WG01 1.03 WG01 56 43 52 WG01,CK79,C90 −3.55 D95
F12 13.62 KH95 12.30 KH95 1.18 KH95 80.8 CK79 −4.19 G07
F13 . . . . . . . . . . . . . . . . . . . . . . . . −3.65 D95
F14 13.93 KH95 12.37 KH95 2.32 KH95 53.5 K98 . . . . . .
F15 12.28 KH95 11.11 KH95 1.25 KH95 7.8 CK79 −3.97 G07
F16 8.62 KH95 8.10 KH95 0.90 KH95 3.5 5 CK79,C90 −2.61 G07
F17 14.30 HH92 12.50 HH92 2.14 HH92 66 46 SA11,CK79 . . . . . .
F18 12.17 KH95 11.19 KH95 0.59 KH95 9 5 B90,C90 −3.54 N95
Outflow Disks
O1 12.33 KH95 11.42 KH95 2.29 KH95 137.9 R10 −3.13 G07
O2 11.07 KH95 9.87 KH95 0.21 KH95 53.9 CK79 −3.78 D95
O3 11.51 KH95 10.54 KH95 3.20 HEG95 112.8 73 110 CK79,E94,C90 −4.39 . . .
O4 . . . . . . . . . . . . . . . . . . . . . . . . −2.60 G07
O5 12.41 KH95 11.37 HHG94 2.64 KH95 108.9 101 CK79,C90 −4.27 B99
O6 13.09 KH95 11.95 KH95 1.46 KH95 85.4 CK79 −4.58 G07
O7 12.15 KH95 11.06 KH95 0.87 KH95 22.5 17 K98,CK79 −3.66 G07
O8 14.85 KGW08 13.54 L00 11.90 K09 88.2 CK79 −3.68 G07
O9 12.96 KH95 12.17 KH95 0.52 KH95 158 E87 −4.40 G07
O10 12.68 KH95 9.87 KH95 1.91 KH95 8.5 E87 −4.46 N95
O11 9.95 KH95 9.34 KH95 0.50 F09 84.2 CK79 −4.03 D95
O12 9.53 KH95 8.80 KH95 1.84 KH95 21 B90 −2.87 G07
O13 9.19 KH95 8.50 KH95 1.39 KH95 38 T09 −2.79 C98
O14 11.92 KH95 10.83 KH95 1.35 KH95 47 72.8 E87,CK79 . . . . . .
Table 2 (Continued)
References.Alcal´a et al.2000(A00); Andrews et al.2011(A11); Antoniucci et al.2011(SA11); Bary et al.2008(B08); Beckwith et al.1990(B90); Bouvier & Appenzeller1992(BA92); Bouvier et al.1993(B93); Brice˜no et al.1999(B99); Cabrit et al.1990(C90); Carkner et al.1998(C98); Chen et al.1995(CMLW95); Cohen & Kuhi1979(CK79); Cutri et al.2003(C03); Damiani et al.1995(D95); DENIS Consortium2005(DENIS); Edwards et al.1987(E87); Edwards et al.1994 (E94); Espaillat et al.2011(E11); Feigelson & Kriss1989(FK89); Feigelson et al.1993(F93); Fernandez1995(F95); Furlan et al.2009(F09); Gauvin & Strom1992 (GS92); G¨udel et al.2007(G07); Guenther & Emerson1997(GE97); Hartigan et al.1995(HEG95); Herbst et al.1994(HHG94); Herczeg et al.2007(H07); Hughes et al.1994(H94); Kastner et al.1999(K99); Kenyon & Hartmann1995(KH95); Kenyon et al.1998(K98); Kenyon et al.2008(KGW08); Kraus & Hillenbrand2009 (K09); Luhman2000(L00); Luhman2004(L04); Magazzu et al.1992(M92); McClure et al.2010(MFM10); Mer´ın et al.2010(M10); Monet et al.2003(M03); Muzerolle et al.2001(MCH01); Neuhaeuser et al.1995(N95); Osterloh & Beckwith1995(OB95); Prato et al.2003(PGS03); Rebull et al.2010(R10); Riaz et al. 2006(R06); Salyk et al.2009(SB09); Schisano et al.2009(S09); Spezzi et al.2007(S07); Taguchi et al.2009(T09); White & Ghez2001(WG01); White et al.2000 (W00); Wilking et al.2005(WMRG05).
R-band extinctions via the relationships of Mathis (1990). The
de-reddened line fluxes were converted to line luminosities us-ing the known distances to each different star-formus-ing region
(see Table1and references therein). The Hαline luminosities,
LHα, were then converted into accretion luminosities,Lacc, with
the empirical relationships of Fang et al. (2009):
log (Lacc/L)=(2.27±0.23) + (1.25±0.007)
×log (LHα/L). (1)
Our derived accretion luminosities are listed in Table3. From our
accretion luminosities, we then used the empirical relationships
of Yang et al. (2012) to relate accretion luminosities to FUV
luminosities:
log (LFUV/L)= −1.670 + 0.836×log (Lacc/L). (2)
Our derived FUV luminosities are also listed in Table3. For
the 12 disks shared between this study and Yang et al. (2012),
we found that our FUV luminosities agreed to those derived by
Yang et al. (2012) within 0.35 dex. We found no systematic shift
between our Hα-derived FUV luminosities and their directly
measured FUV luminosities. Stellar chromospheric activity
can also result in Hα emission, so we used the spectral type
dependent, equivalent width cutoffs of White & Basri (2003) to
distinguish between chromospheric activity and accretion. For
targets where the Hαequivalent width fell below these cutoffs,
we report accretion and FUV luminosity upper limits.
Converting accretion luminosities to accretion rates requires some physical knowledge of the system and the processes of
accretion. Gullbring et al. (1998) developed a simple
magne-tospheric accretion model whereby the accretion luminosity is generated by the release of potential energy as gas falls from the inner edge of the disk onto the surface of the star along
stellar magnetic field lines. In this model, the accretion rate,M˙,
is related to accretion luminosity by
˙
M= LaccR
GM
1− R
Rin
, (3)
where R and M are the radius and mass of the star, G is
Newton’s gravitational constant, andRinis the inner truncation
radius of the disk. Rin is generally unknown, but is usually
assumed to be ≈5R, which corresponds to the typical
co-rotation distance (Gullbring et al.1998; Shu et al.1994). The
stellar radius is determined from the star’s effective temperature and bolometric luminosity via the Stefan–Boltzmann Law. We used pre-main sequence evolutionary tracks from Siess
et al. (2000) to relate the effective temperature and bolometric
luminosity to specific stellar masses. Our final accretion rates
(as well as stellar masses) are listed in Table3.
To test the validity of our self-consistently derived accretion rates, we compared our results with an array of other studies,
including Najita et al. (2007), Gullbring et al. (1998), Hartmann
et al. (1998), White & Ghez (2001), and Hartigan et al.
(1995). While differences between accretion rates can develop
from several factors (including the use of different accretion tracers, different estimates of extinction, different bolometric corrections, use of non-contemporaneous photometry, etc.), we find that our accretion rates generally agree with past studies to
within∼0.5 dex. This level of variation between accretion rates
computed from different tracers is typical, even if observations
are contemporaneous (Rigliaco et al.2012). Our accretion rates
are also not significantly offset from past studies of accretion
rates, with the exception of Hartigan et al. (1995), who find
systematically higher accretion rates (although this systematic offset from other estimates has been noted in previous studies;
e.g., Gullbring et al.1998).
One of the major advantages of our study, as compared to many past studies, is that our accretion rates are self-consistently derived using all the same metric, instead of being aggregated from different literature sources which adopt different methods.
2.3.3. Disk Structure and Disk Mass
Many transitional disks in our sample have been previously modeled with radiative transfer codes in order to reproduce near-and mid-infrared disk spectra near-and resolved millimeter images. While the exact nature of these disk models can vary between papers, they all involve the creation of a simple, axisymmetric model disk with a prescribed dust and gas surface density. Models specific to transitional disks include gas and dust cavities
within a specified radius:rcavity. At the outer edge of this dust
cavity, the frontally illuminated disk wall puffs up to a wall
height ofhwall, which can significantly affect the near-infrared
emission of transitional disks (both due to excess emission
and shadowing of the outer disk; Espaillat et al.2011). These
disk models are then subject to simulated observations, and the relevant model spectra or resolved images are calculated (for some specified viewing angle) and fit to observations. We
aggregated values for the cavity size (rcavity) and the wall height
(hwall) from the literature. These disk properties are listed in
Table4. While the individual models can vary between papers,
the majority of these cavity sizes and wall heights are taken
from Andrews et al. (2011) and Espaillat et al. (2011), which
both use similar disk models.
Table 3
Stellar and Accretion Properties
ID Teff M Lbol R logLacc logM˙ logLFUV
(K) (M) (L) (R) (L) (Myr−1) (L
) Transition Disks
T1 6030 1.15 1.01 0.92 . . . . . . . . .
T2 5520 1.34 1.98 1.54 −2.0 −9.3 −3.3
T3 5770 3.7 47.60 6.92 −1.2 −8.3 −2.7
T4 4900 1.98 2.83 2.34 −0.1 −7.4 −1.8
T5 4205 1.62 3.76 3.66 −1.2 −8.2 −2.6
T6 3705 0.45 0.18 1.02 −1.3 −8.4 −2.8
T7 4350 0.95 0.33 1.01 −1.0 −8.4 −2.5
T8 4730 1.26 0.82 1.35 −0.5 −7.8 −2.1
T9 4277.5 1.00 0.50 1.28 −0.7 −7.9 −2.2
T10 3777.5 0.52 0.46 1.59 −4.0 −11.0 −5.0
T11 4350 1.05 0.53 1.29 −1.5 −8.9 −3.0
T12 5830 1.25 2.78 1.64 −0.4 −7.7 −2.0
T13 4590 1.28 1.01 1.59 −0.9 −8.2 −2.5
T14 3125 0.16 0.08 0.94 −3.4 −10.0 −4.5
T15 3632.5 0.40 0.16 1.00 −1.4 −8.4 −2.8
T16 3487.5 0.35 0.21 1.26 −3.0 −9.9 −4.2
T17 4060 0.80 0.23 0.97 −1.1 −8.4 −2.6
T18 3777.5 0.52 0.35 1.38 −1.3 −8.3 −2.7
T19 3060 0.17 0.17 1.45 −2.5 −8.9 −3.7
T20 3777.5 0.52 0.18 0.99 −1.4 −8.5 −2.8
T21 5250 2.18 5.50 2.84 −1.0 −8.2 −2.5
T22 5520 1.31 0.89 1.04 −1.6 −9.1 −3.0
T23 3850 0.58 0.39 1.40 −1.0 −8.0 −2.5
T24 4900 1.3 1.21 1.53 −2.0 −9.3 −3.3
T25 3197.5 0.17 0.04 0.64 −2.9 −9.7 −4.1
T26 5845 2.20 11.06 3.25 0.3 −6.9 −1.4
Full Disks
F1 4060 0.8 0.43 1.32 −1.3 −8.5 −2.8
F2 4060 0.78 0.70 1.70 −0.9 −8.0 −2.4
F3 4060 0.78 0.67 1.65 −0.6 −7.7 −2.2
F4 3705 0.46 0.32 1.37 −1.4 −8.3 −2.8
F5 3560 0.38 0.59 2.03 −0.9 −7.6 −2.4
F6 4060 0.76 0.78 1.79 −1.4 −8.4 −2.8
F7 4060 0.76 0.80 1.81 −0.3 −7.3 −1.9
F8 3850 0.56 0.70 1.88 −1.7 −8.5 −3.0
F9 3850 0.57 0.42 1.46 −0.8 −7.8 −2.4
F10 4350 1.2 0.46 1.20 −1.2 −8.6 −2.7
F11 4060 0.76 0.90 1.92 −0.7 −7.7 −2.3
F12 3850 0.57 0.18 0.95 −1.5 −8.7 −3.0
F13 . . . . . . . . . . . . . . . . . . . . .
F14 3777.5 0.52 0.28 1.23 −1.5 −8.5 −2.9
F15 3777.5 0.52 0.55 1.74 −2.1 −9.0 −3.4
F16 5860 1.8 7.96 2.74 −0.7 −8.0 −2.3
F17 3415 0.33 0.34 1.67 −1.6 −8.3 −3.0
F18 4060 0.72 0.37 1.23 −2.4 −9.5 −3.6
Outflow Disks
O1 4730 1.07 0.66 1.21 −0.2 −7.5 −1.8
O2 3415 0.33 1.25 3.20 −0.9 −7.3 −2.4
O3 4205 0.9 2.06 2.71 0.4 −6.5 −1.4
O4 3850 . . . . . . . . . . . . . . . . . .
O5 3850 0.56 0.80 2.01 −0.2 −7.1 −1.9
O6 3777.5 0.52 0.28 1.24 −1.1 −8.2 −2.6
O7 4205 0.95 0.46 1.27 −1.7 −9.0 −3.1
O8 3850 0.55 6.42 5.71 1.9 −4.5 −0.1
O9 4350 0.7 0.14 0.67 −1.1 −8.5 −2.6
O10 3705 0.45 2.34 3.72 −2.0 −8.5 −3.4
O11 4730 1.49 2.03 2.13 0.1 −7.2 −1.6
O12 5080 2.2 6.15 3.21 0.0 −7.2 −1.7
O13 5250 2.18 6.77 3.15 0.3 −6.9 −1.4
O14 4060 0.77 0.72 1.72 −0.8 −7.8 −2.3
Table 4 Disk Properties
ID rgap Ref. hwall Ref. f850μm f1.3mm Ref. M850μm M1.3mm
(AU) (AU) (mJy) (mJy) (MJup) (MJup)
Transition Disks
T1 . . . . . . . . . . . . . . . . . . . . .
T2 37.1 KM09 . . . . . . . . . 118 H93 . . . 25.8
T3 146.7 KM09 . . . . . . . . . 143 H93 . . . 32.5
T4 10 E11 . . . . . . . . . 124.9 H93 . . . 28.4
T5 38 E11 7 E11 . . . 128.4 H93 . . . 29.1
T6 19 A11 5.7 A11 237 109 M13 13.6 19.0
T7 . . . . . . . . . . . . . . . 75 M13 . . . 9.6
T8 30 A11 9 A11 181 105 M13 7.6 13.4
T9 23 E11 2.9 E11 . . . 253 M13 . . . 44.1
T10 . . . . . . . . . . . . . . . . . . . . . . . . . . .
T11 39 E11 5 E11 428 167 M13 24.5 29.1
T12 68 A11 6.8 A11 . . . 70 OB95 . . . 38.9
T13 30 A11 2 A11 . . . . . . . . . . . . . . .
T14 . . . . . . . . . . . . . . . . . . . . . . . . . . .
T15 . . . . . . . . . . . . . . . . . . . . . . . . . . .
T16 13 E11 2 E11 . . . 105 H93 . . . 23.9
T17 5 E11 4 E11 . . . 100 H93 . . . 22.8
T18 20 E11 4 E11 . . . 47.8 H93 . . . 10.9
T19 55 M10 . . . . . . . . . 36 N97 . . . 7.7
T20 . . . . . . . . . . . . . . . 27 N97 . . . 5.8
T21 18 E11 4 E11 . . . 77.5 H93 . . . 17.7
T22 15 BR07 . . . . . . . . . 105.2 H93 . . . 11.1
T23 4 T10 . . . . . . . . . . . . . . . . . . . . .
T24 . . . . . . . . . . . . 173 63 M13 3.6 11.0
T25 15 A11 0.8 A11 149 89 M13 6.3 11.4
T26 36 A11 8.2 A11 397 95 M13 16.7 12.2
Full Disks
F1 . . . . . . . . . . . . 144 88 M13 8.3 15.3
F2 . . . . . . . . . . . . 130 47 M13 7.5 8.2
F3 . . . . . . . . . . . . 324 190 M13 18.6 33.1
F4 . . . . . . . . . . . . . . . 111 G11 19.3
F5 . . . . . . . . . . . . 90 36 M13 5.2 6.3
F6 . . . . . . . . . . . . 80 35 M13 4.6 6.1
F7 . . . . . . . . . . . . 440 230 M13 25.2 40.1
F8 . . . . . . . . . . . . 201 84 M13 11.5 14.6
F9 . . . . . . . . . . . . 208 91 M13 11.9 15.9
F10 . . . . . . . . . . . . . . . . . . . . . . . . . . .
F11 . . . . . . . . . . . . 1255 593 M13 72.0 103.3
F12 . . . . . . . . . . . . 173 83 M13 9.9 14.5
F13 . . . . . . . . . . . . . . . . . . . . . . . . . . .
F14 . . . . . . . . . . . . . . . . . . . . . . . . . . .
F15 . . . . . . . . . . . . 178 87 M13 5.0 15.2
F16 . . . . . . . . . . . . 74 30 M13 4.2 5.2
F17 . . . . . . . . . . . . . . . . . . . . . . . . . . .
F18 . . . . . . . . . . . . 74 37 M13 4.2 6.4
Outflow Disks
O1 . . . . . . . . . . . . 66 96 M13 3.8 16.7
O2 . . . . . . . . . . . . 8.8 25 M13 0.5 4.4
O3 . . . . . . . . . . . . . . . 389.9 G11 . . . 67.9
O4 . . . . . . . . . . . . . . . . . . . . . . . . . . .
O5 . . . . . . . . . . . . 248 136 M13 14.2 23.7
O6 . . . . . . . . . . . . 10 27 M13 0.6 4.7
O7 . . . . . . . . . . . . 33 21 M13 1.9 3.7
O8 . . . . . . . . . . . . . . . 34.2 G11 . . . 5.9
O9 . . . . . . . . . . . . 29 15 M13 1.7 2.6
O10 . . . . . . . . . . . . 47 40 A05 2.7 7.0
O11 . . . . . . . . . . . . 79 42 M13 4.5 7.3
O12 . . . . . . . . . . . . 560 229 M13 32.1 39.9
O13 . . . . . . . . . . . . 628 280 M13 36.0 48.8
O14 . . . . . . . . . . . . 102 29 M13 5.9 5.1
O15 . . . . . . . . . . . . 560 172 M13 32.1 30.0
mass, from 1.3 mm and 850μm photometry available in the
literature. Following Beckwith et al. (1990), it is possible to
invert observed millimeter flux into an apparent disk dust mass if we assume that the emission is (1) optically thin, (2) arises from an isothermal region of the disk of known temperature, and (3) is due to material with a known opacity. (See Beckwith et al.
1990and Mohanty et al.2013for a more detailed explanation
of this process, and the assumed dust temperatures and dust opacities.) Using the canonical gas-to-dust ratio of 100-to-1, we then converted dust masses into total gas masses. The resulting
total disk masses are listed in Table4. It is important to note
that even if we disregard uncertainties in the dust temperature or opacity and the questionable gas-to-dust ratio, these disk masses are likely lower limits. Millimeter observations are
only sensitive to small dust grains, less than ∼1 cm in size.
It is possible that substantial mass may be in larger grains, planetesimals, or even protoplanets.
3. RESULTS: DETECTION OF [OI] 63.18μm AND
O-H2O EMISSION
We detect [Oi] 63.18μm emission from 17 of our 21
transitional disks. Coupling these new results with our reanalysis
of select disks from the GASPS sample (Dent et al. 2013;
Howard et al.2013), we report [Oi] 63.18μm emission from
21 of 26 transitional disks, 12 of 18 full disks, and emission from all of the outflow disks. We fit all observed emission lines to Gaussians using an original MATLAB fitting routine. To mitigate noise in the PACS spectrum, we fit the lines over a range of wavelength baselines (the minimum wavelength range:
63.13−63.23μm; the maximum wavelength range spanned
the entire PACS spectrum: 62.93−63.43μm). The best-fitting
spectrum was deemed as the spectrum closest to the median of all line fits for a given target. The line flux of this best-fitting spectrum was calculated from the (continuum-subtracted)
Gaussian line profile (flux = amplitude·σGaussian
√
2π). For
[Oi] 63.18μm non-detections, we derive 3σ upper limits
assuming a Gaussian profile with a 3σrms peak height (where
σrms is the standard deviation of the continuum linear-fit)
and a 98 km s−1 line width corresponding to the FWHM
of an unresolved line in PACS (PACS Observer’s Manual).
Continuum fluxes at 63μm were also found from the
best-fitting Gaussian line profile, as the constant baseline flux term.
Continuum emission at 63μm was detected for all targets,
with the exception of DS Tau (an upper limit of 0.037 Jy).
The [Oi] 63μm line fluxes and 63μm continuum fluxes are
reported in Table5, and the spectra are provided in theAppendix.
To validate our data reduction, we compared our resulting
[Oi] 63μm line fluxes and 63μm continuum fluxes to the fluxes
reported by Howard et al. (2013). Despite using a more recent
version of HIPE (version 9, rather than version 4), our fluxes
generally agree with those of Howard et al. (2013) to within
∼30%, which is comparable to the absolute flux accuracy of
PACS (which has a peak-to-peak accuracy 30%, and rms
accuracy of10%;PACS Observer’s Manual). Compared to
this pipeline uncertainty, the uncertainties in our line fits are negligible. Representative error bars for both of these types of
flux calibration uncertainty are shown in Figures1(a) and (b).
Typical line fluxes (normalized to the distance of the Taurus-Auriga star-forming region, at 140 pc) are on the
or-der of 10−16–10−17 W m−2, corresponding to line
luminos-ity of 10−7–10−5L
. Continuum fluxes (again, normalized to
140 pc) range from 0.1–100 Jy, corresponding to continuum
luminosities (Lcontinuum=fνν4π d2) of 10−2–1L. Figure1(a)
shows the [Oi] 63.18μm line luminosity as a function of 63μm
continuum luminosity for all of our targets. Figure1(b) shows
the ratio of [Oi] 63.18μm line luminosity to 63μm continuum
luminosity as a function of 63μm continuum luminosity for all
of our targets.
We used the Astronomy SURVival package (ASURV;
LaValley et al. 1992) to perform linear regressions and
cor-relation tests between the line and continuum luminosities for each subsample. ASURV is particularly useful as it allows for the incorporation of censored data points (i.e., non-detection, line flux upper limits). Compared to the other subsamples, we oversample G-type stars in transitional disks (five G-type tran-sitional disks; one G-type full disk; zero G-type outflow disks). Because of this oversampling and the seemingly chaotic nature of the G-type line and continuum fluxes, we have omitted them
from many of our statistical tests.11Additionally, as discussed in
Section2.1, we also exclude multiple systems where the
multi-plicity likely strongly affects ourHerschel/PACS observations.
Tables 6–8 summarize the results from a variety of statistics
and fitting routines that were used to characterize differences between the three subsamples. There are a number of
impor-tant trends in our [Oi] 63.18μm line and 63μm continuum
luminosity data.
1. [Oi] 63.18μm line luminosities are positively correlated
with 63μm continuum luminosities, both for the sample
as a whole and for each individual subsample, as shown
in Table 6. This correlation was previously recognized
in the Herschel/PACS GASPS survey of Taurus-Auriga
protoplanetary disks (Howard et al. 2013) and Herbig
Ae/Be stars (Meeus et al.2012), though our study extends
this result to a significantly larger sample of transitional disks.
2. Outflow disks tend to have [Oi] 63.18μm line
lumi-nosities and 63.18μm continuum luminosities that differ
markedly from full disks and transitional disks. This is
simply demonstrated in Table7, which shows that both the
line and continuum luminosities of outflow disks are not likely from the same parent population as either the full
disks or transitional disks. As shown in Table 8, outflow
disks tend to have higher [Oi] 63.18μm line luminosities
(by 0.5–1 dex), higher 63μm continuum luminosities (by
0.5 dex), and higher line-to-continuum luminosity ratios
(by0.5 dex). This was previously recognized by Podio
et al. (2012) and Howard et al. (2013).
3. Full disks and transitional disks have similar 63μm
con-tinuum luminosities. This is most easily shown in Table7,
which shows that the 63μm continuum luminosities of
tran-sitional disks and full disks are effectively indistinguish-able.
4. Given the same 63.18μm continuum luminosity, full disks
tend to have larger [Oi] 63.18μm line luminosities than
transitional disks,12 by a factor of2. While this is
visu-ally evident in Figure1(a), there is sufficient scatter (and
non-detections) to make this difficult to quantify, and the
11 Our oversampling of G-type transitional disks is not intentional. Due to the
rarity of transitional disks, we cannot discriminate transitional disks by spectral type in order to populate our transitional disk subsample.
Simultaneously, it is difficult to populate subsamples of full or outflow disks with G-type stars from the GASPS Taurus-Auriga survey, since Taurus-Auriga is a low-mass star-forming region.
12 The one notable exception is BP Tau (C2). BP Tau has a significantly lower
Table 5 HerschelPACS Results
ID [Oi] 63.18μm Line Flux o-H2O 63.32μm Line Flux 63μm Continuum Flux
(10−17W m−2) (10−17W m−2) (Jy)
Transition Disks
T1 2.340±0.143 0.716 3.029±0.010
T2 4.492±0.200 1.192 0.596±0.020
T3 11.284±0.851 5.325 2.262±0.055
T4 1.530±0.150 0.658 1.618±0.015
T5 2.200±0.134 0.702 4.017±0.010
T6 1.276±0.305 0.960 0.972±0.016
T7 0.949±0.076 0.541 0.829±0.013
T8 2.536±0.189 1.014±0.145 5.315±0.012
T9 3.793±0.513 2.435 2.810±0.036
T10 0.764±0.101 0.523 0.195±0.008
T11 1.306±0.165 0.770 1.285±0.012
T12 1.276 1.276 13.239±0.019
T13 1.934±0.118 0.534 1.317±0.009
T14 1.078 1.078 0.303±0.016
T15 0.910±0.112 0.540 1.406±0.008
T16 0.823±0.140 0.817 0.652±0.012
T17 1.390±0.124 0.755 0.556±0.011
T18 0.354±0.097 0.637 0.810±0.008
T19 0.770 0.770 0.490±0.013
T20 1.035±0.112 0.730 0.779±0.016
T21 1.777±0.100 0.793 4.199±0.008
T22 5.455±0.291 1.493 7.318±0.021
T23 4.239±0.345 1.556 3.675±0.024
T24 3.894±0.266 1.671 4.021±0.024
T25 0.590 0.590 0.690±0.009
T26 1.543±0.296 2.534 39.540±0.023
Full Disks
F1 2.606±0.109 0.956±0.131 1.106±0.016
F2 0.647±0.146 0.898±0.173 0.490±0.024
F3 2.016±0.183 0.921 0.979±0.014
F4 1.489±0.416 1.573 0.126±0.027
F5 0.712±0.384 2.150 1.946±0.039
F6 1.748±0.155 0.455±0.157 0.932±0.012
F7 2.792±0.208 0.640±0.176 1.268±0.012
F8 1.023 1.023 0.780±0.015
F9 2.505±0.351 1.413 1.292±0.021
F10 0.782 0.782 0.037
F11 6.185±0.381 1.850 3.756±0.028
F12 0.969 0.969 0.343±0.019
F13 1.585 1.585 0.086±0.024
F14 3.847±0.262 1.234 2.428±0.023
F15 1.512±0.256 0.968±0.294 0.744±0.019
F16 12.650±0.336 1.474 9.043±0.027
F17 0.866 0.866 0.719±0.014
F18 1.259 1.259 0.370±0.015
Outflow Disks
O1 9.061±0.364 1.601 1.707±0.024
O2 4.541±0.333 1.410 0.369±0.035
O3 187.160±3.075 10.125 18.015±0.224
O4 80.446±1.891 6.355 13.600±0.113
O5 39.268±1.971 7.357 3.932±0.129
O6 4.069±1.085 3.880 0.411±0.060
O7 2.592±0.803 4.039 1.286±0.057
O8 8.075±0.397 2.109 6.240±0.030
O9 5.272±0.221 1.130 1.020±0.017
O10 10.000±0.548 2.385 1.376±0.034
O11 17.452±0.701 0.947±0.320 2.109±0.043
O12 13.020±0.640 2.481±0.740 14.103±0.049
O13 1348.400±14.564 46.876±1.506 161.870±0.966
O14 38.197±0.828 1.660±0.389 6.650±0.058
O15 2.573±0.554 2.789 1.062±0.046
Table 6 Subsample Correlation Tests
Correlation Subsample Correlation Tests Correlated? Linear Regression
Test Being Tested P(1) P(2) P(3) Intercept Slope
L 63μm v. L [Oi] 63μm All Objects 0.0% 0.0% 0.0% Correlated −3.24±0.18 1.15±0.13
Transitional Disks Only 0.0% 0.0% 0.1% Correlated −4.07±0.21 0.74±0.15
Full Disks Only 10.0% 0.7% 3.3% Correlated −4.42±0.31 0.38±0.18
Outflow Disks Only 0.2% 0.2% 0.5% Correlated −3.01±0.17 0.97±0.15
Teffv. L [Oi] 63μm All Objects 0.1% 0.3% 0.4% Correlated −7.66±0.93 0.0007±0.0002
Transitional Disks Only 0.0% 0.1% 0.2% Correlated −6.67±0.39 0.0004±0.0001
Full Disks Only 82.7% 84.7% 87.3% Not Correlated −4.68±1.15 −0.0001±0.0003
Outflow Disks Only 24.7% 29.7% 33.0% Not Correlated −6.22±1.70 0.0005±0.0004
Teffv. L 63μm All Objects 10.9% 0.4% 1% Not Correlated −3.09±0.73 0.0004±0.0002
Transitional Disks Only 0.0% 0.1% 0.2% Correlated −3.17±0.30 0.0004±0.0001
Full Disks Only 10.4% 28.4% 18.0% Not Correlated 1.81±1.63 −0.0009±0.0004
Outflow Disks Only 11.8% 14.4% 15.6% Not Correlated −3.79±1.45 0.0007±0.0003
Lbolv. L [Oi] 63μm All Objects 0.0% 0.0% 0.0% Correlated −4.61±0.10 1.02±0.19
Transitional Disks Only 0.1% 0.3% 0.3% Correlated −4.96±0.07 0.41±0.11
Full Disks Only 28.3% 45.7% 38.3% Not Correlated −4.94±0.19 0.39±0.50
Outflow Disks Only 3.0% 7.3% 4.7% Correlated −4.04±0.20 0.75±0.35
Lbolv. L 63μm All Objects 0.0% 0.0% 0.0% Correlated −1.23±0.09 0.78±0.15
Transitional Disks Only 0.0% 0.1% 0.2% Correlated −1.23±0.06 0.52±0.09
Full Disks Only 29.0% 32.9% 36.9% Not Correlated −1.60±0.39 0.64±0.99
Outflow Disks Only 0.2% 0.4% 0.8% Correlated −1.08±0.14 0.97±0.25
LXv. L [Oi] 63μm All Objects 5.9% 59.0% 80.7% Not Correlated −5.59±0.50 −0.18±0.13
Transitional Disks Only 6.4% 56.9% 53.7% Not Correlated −5.62±0.29 −0.14±0.08
Full Disks Only 65.5% 92.0% . . . Not Correlated −5.35±0.79 −0.06±0.20
Outflow Disks Only 9.1% 39.5% 54.8% Not Correlated −2.69±0.64 0.40±0.18
LXv. L 63μm All Objects 3.0% 69.6% 13.9% Not Correlated −2.21±0.36 −0.18±0.10
Transitional Disks Only 0.3% 37.1% 27.6% Not Correlated −2.32±0.34 −0.24±0.09
Full Disks Only 86.4% 73.2% 80.9% Not Correlated −2.03±0.58 −0.03±0.15
Outflow Disks Only 0.8% 8.9% 18.5% Not Correlated 0.59±0.53 0.47±0.15
Laccv. L [Oi] 63μm All Objects 0.0% 0.0% 0.0% Correlated −4.22±0.15 0.66±0.12
Transitional Disks Only 1.0% 4.7% 4.9% Correlated −4.76±0.15 0.28±0.11
Full Disks Only 3.5% 17.6% 16.1% Not Correlated −4.69±0.27 0.34±0.22
Outflow Disks Only 5.7% 5.2% 8.6% Not Correlated −3.88±0.23 0.41±0.22
Laccv. L 63μm All Objects 0.0% 0.0% 0.0% Correlated −0.97±0.14 0.47±0.10
Transitional Disks Only 0.3% 4.1% 2.8% Correlated −1.06±0.14 0.29±0.09
Full Disks Only 3.6% 19.9% 18.2% Not correlated −1.21±0.61 0.54±0.44
Outflow Disks Only 0.7% 1.6% 1.2% Correlated −0.85±0.19 0.48±0.18
˙
Mv. L [Oi] 63μm All Objects 0.0% 0.1% 0.1% Correlated −0.44±0.94 0.55±0.12
Transitional Disks Only 4.0% 13.5% 12.1% Not Correlated −3.03±1.15 0.25±0.14
Full Disks Only 7.0% 36.8% 25.2% Not Correlated −2.68±1.66 0.30±0.20
Outflow Disks Only 10.8% 1.5% 4.5% Correlated −1.91±1.44 0.28±0.19
˙
Mv. L 63μm All Objects 0.0% 0.0% 0.0% Correlated 2.12±0.81 0.44±0.10
Transitional Disks Only 0.6% 1.2% 2.5% Correlated 1.24±1.01 0.32±0.12
Full Disks Only 0.7% 4.4% 4.3% Correlated 4.50±3.29 0.77±0.40
Outflow Disks Only 1.2% 1.0% 1.0% Correlated 1.56±1.21 0.35±0.16
mdiskv. L [Oi] 63μm All Objects 0.0% 26.5% 43.6% Not Correlated −5.79±0.25 0.04±0.01
Transitional Disks Only 0.3% 26.6% 20.0% Not Correlated −4.72±0.18 0.03±0.01
Full Disks Only 1.9% 44.7% 8.2% Not Correlated −5.25±0.13 0.01±0.01
Outflow Disks Only 0.0% 4.4% 2.3% Correlated −4.78±0.26 0.03±0.01
mdiskv. L 63μm All Objects 0.0% 5.7% 7.9% Correlated −2.14±0.16 0.03±0.01
Transitional Disks Only 0.6% 7.7% 5.6% Correlated −2.13±0.25 0.03±0.01
Full Disks Only 6.6% 95.1% 86.7% Not Correlated −2.05±0.20 0.02±0.01
Outflow Disks Only 0.0% 4.4% 13.0% Correlated −1.71±0.24 0.03±0.01
acavityv. L [Oi] 63μm Transitional Disks Only 71.1% 77.0% 69.5% Not Correlated −5.10±0.17 −0.0012±0.01
acavityv. L 63μm Transitional Disks Only 75.0% 24.6% 26.5% Not Correlated −1.46±0.16 0.00±0.01
hwallv. L [Oi] 63μm Transitional Disks Only 23.6% 78.5% 64.7% Not Correlated −5.25±0.20 0.04±0.04
hwallv. L 63μm Transitional Disks Only 31.0% 23.6% 24.5% Not Correlated −1.65±0.18 0.07±0.04
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 −5.5
−5 −4.5 −4 −3.5 −3 −2.5 −2
log 63 µm continuum luminosity, L
Sun
log [OI] 63
µm line luminosity, L
Sun
Outflow Disks
Full Disks
Transition Disks
30% Peak−to−Peak Absolute Calibration Accuracy
10% RMS Absolute Flux Calibration Accuracy 10% Relative Flux Calibration Accuracy
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5
−17.5 −17 −16.5 −16 −15.5 −15 −14.5 −14 log 63 µm continuum flux, at 140 pc, Jy
log [OI] 63
µm line flux, at 140 pc, W/m
2
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1
−5.5 −5 −4.5 −4 −3.5 −3 −2.5 −2 −1.5 −1
log 63 µm continuum luminosity, LSun
log [OI] 63
µm line luminosity, L
Sun
/ 63 µm continuum luminosity, L
Sun
Outflow Disks Full Disks Transition Disks (a)
[image:12.612.149.465.58.549.2](b)
Figure 1.(a) [Oi] 63.18μm line luminosity as a function of 63μm continuum luminosity for our sample of transitional disks (red), full disks (blue), and outflow disks (green). upper limits of 3σare denoted by hollow data points with arrows. Symbols correspond to stellar spectral types: circles are G-type stars (which are included in this plot but neglected in the statistical analysis, for reasons described in the paper), squares are K-type stars, and diamonds are M-type stars. BP Tau (an evolved full disk) is indicated in purple. Targets excluded from statistical tests (for either being a binary that does not meet the criteria in Section2.1, or being a G-type star) are marked by an asterisk. (b) The ratio of [Oi] 63.18μm line luminosity/63μm continuum luminosity as a function of 63μm continuum luminosity for our sample of transitional disks (red), full disks (blue), and outflow disks (green). upper limits of 3σare denoted by hollow data points with arrows. Symbols are as in panel (a). (A color version of this figure is available in the online journal.)
ASURV statistical tests point to indistinguishable line
lumi-nosities between the two subsamples (see Tables7and8).
However, this difference between full disks and transitional disks becomes clear when we examine the ratio of the
[Oi] 63.18μm line luminosity to the 63μm continuum
luminosity, as shown in Figure1(b). The ASURV
statisti-cal tests indicate that the distribution of line-to-continuum ratios of transitional disks is significantly different from
that of full disks (see Tables7and8), with full disks having
line-to-continuum ratios larger by a factor of2.
Additionally, the best-fit linear regressions in the line luminosity for full disks and transitional disks, as shown
in Table 6, are distinct. Transitional disks have steeper
best-fit slopes and shallower best-fit intercepts than full disks; both of these effects contribute to larger differences in line luminosity at the relevant continuum luminosities. We checked our ASURV fit results with an alternative
Bayesian metric (linmix_err.pro; Kelly 2007) and found
Table 7
Subsample Statistical Difference Tests
Parameter Subsamples Statistical Difference Tests Different?
Being Compared P(1) P(2) P(3) P(4) P(5)
L [Oi] 63μm Transitional vs. Full Disks 98.4% 98.4% 78.7% 93.8% 94.1% Not Different
Transitional vs. Outflow Disks 0.0% 0.0% 0.0% 0.0% 0.0% Different
Full vs. Outflow Disks 0.0% 0.0% 0.0% 0.0% 0.0% Different
L 63μm Transitional vs. Full Disks 9.8% 10.2% 4.8% 9.8% 10.0% Not Different
Transitional vs. Outflow Disks 2.1% 1.4% 5.9% 5.9% . . . Different
Full vs. Outflow Disks 0.1% 0.0% 0.0% 0.1% 0.0% Different
L [Oi] 63μm/ Transitional vs. Full Disks 2.2% 1.0% 3.0% 3.1% 1.5% Different
L 63μm Transitional vs. Outflow Disks 0.0% 0.0% 0.0% 0.0% 0.0% Different
Full vs. Outflow Disks 1.8% 1.7% 4.0% 2.3% 2.0% Different
Teff Transitional vs. Full Disks 78.9% 79.1% 50.1% 50.1% . . . Not Different
Transitional vs. Outflow Disks 30.8% 31.8% 45.2% 45.2% . . . Not Different
Full vs. Outflow Disks 11.4% 11.5% 10.3% 10.3% . . . Not Different
Lbol Transitional vs. Full Disks 33.7% 34.3% 76.4% 76.4% . . . Not Different
Transitional vs. Outflow Disks 2.5% 2.8% 1.7% 1.7% . . . Different
Full vs. Outflow Disks 3.7% 3.4% 0.5% 0.5% . . . Different
LX Transitional vs. Full Disks 94.3% 94.4% 68.0% 99.0% 1.9% Not Different
Transitional vs. Outflow Disks 74.6% 75.1% 97.1% 73.5% 74.9% Not Different
Full vs. Outflow Disks 94.8% 94.8% 48.9% 94.7% 96.3% Not Different
Lacc Transitional vs. Full Disks 68.8% 69.1% 96.2% 68.9% 69.1% Not Different
Transitional vs. Outflow Disks 1.9% 1.0% 5.4% 1.9% 1.3% Different
Full vs. Outflow Disks 1.4% 0.9% 8.5% 1.4% 0.9% Different
˙
M Transitional vs. Full Disks 100.0% 100.0% 90.0% 99.7% 99.7% Not Different
Transitional vs. Outflow Disks 1.7% 0.8% 4.1% 1.6% 1.0% Different
Full vs. Outflow Disks 1.9% 1.3% 5.2% 2.0% 1.5% Different
mdisk Transitional vs. Full Disks 59.6% 59.8% 42.6% 53.6% 54.5% Not Different
Transitional vs. Outflow Disks 50.7% 50.5% 50.7% 48.3% 48.7% Different
Full vs. Outflow Disks 75.4% 75.4% 99.8% 75.4% 75.3% Different
Notes.Pis the probability that the parameter being compared between two subsamples is drawn from the same parent distribution; lowPvalues indicate that two subsamples are different. The different statistical tests are (1) Gehan generalized Wilcoxon test (with permutation variance); (2) Gehan generalized Wilcoxon test (with hypergeometric variance); (3) logrank test; (4) Peto & Peto generalized Wilcoxon test; (5) Peto & Prentice generalized Wilcoxon test. If the average of the five statistical tests is less than 5%, they are listed as “different,” in boldface.
This difference between full disks and transitional disks
was previously recognized by Howard et al. (2013) for
the GASPS Taurus-Auriga sample only. Our data extends this trend to a much larger sample of transitional disks,
suggesting that this lower [Oi] 63.18μm line emission is a
characteristic property of transitional disks.
5. There is a weak trend for M-type stars to have lower line and continuum luminosities than K-type stars. This trend is most evident in our sample of transitional disks.
Generally, the [Oi] 63.18μm line is spectrally unresolved.
Most FWHM are within11 km s−1(the native resolution of
PACS) of the expected line width for an unresolved line for
PACS (98 km s−1at 60μm). This result was expected, since
[Oi] emission originates far out in the disk, AU from the
central star (Woitke et al.2010). For gas orbiting a Sun-like star,
Keplerian velocities go asVKeplerian =30 km s−1·(a/AU)−1/2.
Thus, beyond AU from the central star, we expect line
widths on the order of 10’s of km s−1. Even at these
distances, Keplerian velocities dominate over thermal velocities
(∼1 km s−1, assuming typical Oi63.18μm gas temperatures of
∼100 K; Aresu et al.2012), or turbulent velocities (∼1 km s−1;
Hughes et al. 2011), and are the cause of most of the line
broadening. A few objects, all outflow sources, have broader
line widths, as high as 170 km s−1 (e.g., RW Aur). In these
sources, [Oi] 63.18μm emission is thought to originate from
shocks along the jet and/or UV-heated gas in the outflow cavity
walls (Podio et al.2012). Line widths of100’s of km s−1reflect
the similarly large shock velocities. Outflow disks can also have
spatially extended [Oi] 63.18μm emission associated with the
jet, which is detectable in non-central PACS spaxels (Podio et al.
2012). We verified that [Oi] 63.18μm emission was localized
only in the central spaxel for our transitional and full disks. For
outflow disks, we only report [Oi] 63.18μm line and 63μm
continuum fluxes from the central spaxel (for more accurate line and continuum fluxes of outflow disks, including neighboring
spaxels, see Podio et al.2012).
Our PACS spectral range fortuitously also includes the
con-siderably fainter o-H2O 63.32μm emission line. We confirmed
the detection o-H2O emission in five full disks and outflow
disks, previously identified by Riviere-Marichalar et al. (2012).
In addition, we report the marginal detections of o-H2O in IQ
Tau, DK Tau, and BP Tau—for which Riviere-Marichalar et al.
(2012) previously identified 3σ upper limits. These new
detec-tions, from the same observational data, are made possible with our updated version of the Herschel HIPE pipeline and a differ-ent line-fitting algorithm. In addition to these objects, we also